AWS EC2 GPU instances vs Aquanode
Hyperscaler GPU instances (P5, P4d) inside a full enterprise cloud
If you already run on AWS, the GPU instances are next to your VPC, your IAM and your data, and that integration is worth real money. What you pay for it is the per-GPU rate and the fact that an EBS-backed environment is an AWS environment — it does not restore anywhere else.
Where AWS EC2 GPU instances wins
- Everything else in the account: VPC, IAM, S3 adjacency, compliance posture, committed-use and Savings Plan discounts, and a procurement path enterprises already have.
- EFA networking and cluster placement for large distributed training.
- Capacity reservations that actually hold capacity, which no marketplace can promise.
Where Aquanode wins
- Your environment survives the box. Pause a deployment, come back tomorrow, resume with packages, model weights and config intact — instead of rebuilding from a fresh image every session.
- A snapshot taken on one provider restores onto another. Supply moved, price moved, or a region ran dry — the environment follows, rather than being stranded in the account that created it.
- One account across every provider we broker, so you are not maintaining separate logins, keys and billing per vendor to chase capacity.
- Automated backups are on by default when you attach storage at deploy time, at a 6-hour interval — you do not have to remember to configure them.
AWS EC2 GPU instances pricing vs Aquanode
AWS EC2 GPU instances figures are EC2 Capacity Blocks for ML — RESERVED effective hourly rate, US regions. This is NOT the on-demand rate, which is higher; we could not read AWS's on-demand GPU rate from a first-party static page, so we quote only what AWS publishes here., read from their own pricing page on 2026-08-05. The Aquanode column is the lowest live per-GPU rate in our marketplace feed and moves on its own — the two columns are not measured the same way, so treat this as a starting point, not a quote.
Source: https://aws.amazon.com/ec2/capacityblocks/pricing/ — EC2 Capacity Blocks for ML — RESERVED effective hourly rate, US regions. This is NOT the on-demand rate, which is higher; we could not read AWS's on-demand GPU rate from a first-party static page, so we quote only what AWS publishes here.. Verified 2026-08-05. Vendors change prices; check theirs before deciding.
What we do not claim
Restoring an environment requires a snapshot that already exists. Stopping a deployment yourself captures it on the way out, so you can bring it back later on any provider. A provider-side termination is different: it is only recoverable if you had already switched automated snapshots on for that deployment, and it costs you the work since the last one. Automated snapshots are opt-in — nothing runs until you start it — and with none running there is nothing to restore.
`aq deploy --snapshot <id>` rents the cheapest matching GPU and restores your snapshot onto it — including onto a different provider than the one it came from. Creating the snapshot is a separate step today: the standalone ogre CLI on the box writes it.
AWS EC2 GPU instances vs Aquanode: common questions
Are these AWS on-demand prices?
No, and the distinction matters. These are Capacity Blocks for ML reserved rates published by AWS, which are lower than on-demand. AWS's on-demand GPU pricing is served from an interactive calculator rather than a static page, so we do not quote it here rather than sourcing it from a third party.
Should I leave AWS for GPU work?
Not if your data, network and compliance boundary live there — the integration usually outweighs the rate. The case for moving is a workload whose expensive part is the environment, run on boxes that do not need to sit inside your VPC.
